 |
| Figure 1: Orthomosaic Map with GCP Data and Positions |
Introduction:
In this lab we took a more in dept look at ground control points (GCPs). As we discussed in a few previous labs, GCPs are used to correct image data for increased location and elevation accuracy. The main way they accomplish this is by providing a fixed point of reference with know position data that image processing software like Pix4D can use to correct image data. The way it works is similar to a GPS systems use of "checkpoints"; GCPs are basically the same thing except were checkpoints can only provide fixed position data, GCPs are used to correct image data using the position data they provide. You can see what GCPs are and the data they provide in the figure above.
Methods:
For the lab we were given a data set that contained GCP data and used Pix4D to process the data and generate point cloud and pyramid data like we discussed in a previous lab. In figure 2 below you can see the initial quality report generated by Pix4D after initial processing. Note that it contains the GCP data but it gives a red indication beside it, This tells us that the reference data given by the GCPs and the position data found when Pix4D processed the image data do not match within tolerance and must be corrected.
 |
| Figure 2: Initial Quality Report showing GCP Data |
In order to correct the discrepancy between the fixed position data given by the GCPs an the image data generated by Pix4D we use the GCP tools in Pix4D to align the data in the orthomosaic map generated by Pix4D with the GCP markers in the image. Sense Pix4D knows the exact location where the GCPs are once we point out where the GCPs are on the Image data, Pix4D can correct the image data to be more accurate and precise while also giving better image quality.
 |
| Figure 3: GCP Location shown in Ray CLoud. |
In the Figure above you can see the process used to correct the GCP information. Note the blue circles that appear on the surface of the point cloud. These are the locations of the GCPs and the windowed view on the right hand side of the screen shows the area where Pix4D believes that image data to belong. Note that in this figure they do not match; the process of correcting the GCPs involves finding the GCPs on the map and making the image match. This can be a very time consuming process, but in this example we only have to locate 6 GCPs.
Discussion:
 |
| Figure 4: Map Showing GCP Corrections |
 |
| Figure 5: Orthomosaic without GCP Corrections |
Comparing the two figures above, you can see that the overall image quality of the map with GCP corrections is better than the one without. This is due mostly because with the GCP corrections, the program doesnt have to make as many approximations when generating the point cloud and pyramid data, this results in a clearer, sharper image. But, perhaps more importantly for professional applications, the data is now more representative of the real world location in terms of position, elevation, and location.
Conclusions:
GCPs are an essential part of creating accurate image data using GIS. While genrating clearer images is good for aesthetics, the real value in this is gaining more a more accurate representation of the real area being represented. As data without GCPs may be off by several hundred feet or more.
